Numerical Study on Atmospheric, Hydrological and ...

【Research Note】 159

Numerical Study on Atmospheric, Hydrological and Coastal Circulation in Coastal City, Semarang Indonesia

Adi PRASETYO Graduate School for International Development and Cooperation, Hiroshima University, 1-5-1 Kagamiyama, Higashi-Hiroshima 739-8529, Japan Research Centre for Water Resources, Ministry of Public Works Jl. Ir. H. Juanda No. 193, Bandung 40135, Indonesia e-mail: [email protected]

Takao YAMASHITA Graduate School for International Development and Cooperation, Hiroshima University, 1-5-1 Kagamiyama, Higashi-Hiroshima 739-8529, Japan e-mail: [email protected]

Abstract This study is carried out seeing that it is important to understand the interacting components of the Earth system and global climate changes including seasonal changes and its variability in Indonesia. Unfortunately, until now there are only few researches into numerical study on atmospheric, hydrological and coastal circulation topics in coastal city in Indonesia, especially researches which consider the dynamic interactions among the major components of the Earth system, as atmosphere, watershed and the estuary. The objective of this study is to evaluate and validate the applicability of non-hydrostatic meteorological model (MM5), watershed simulation model (HSPF), and three-dimensional hydrodynamic with sediment transport model for estuary and coastal ocean (ECOMSED) in Semarang, Indonesia as part of the Regional Environment Simulator (RES). The simulated results showed that the modeling approach appeared to be a powerful tool to define relevancies between each component in the Earth system. For atmospheric simulation, although some biases were found by comparing the value of observed data and simulated results, trend of precipitation that occured at these meteorological stations shows a similar pattern. In hydrologic modeling, performance of simulation has reflected a good computation by numerical model and results met a good criteria with correlation coefficient (R) of 0.89, determination coefficient (R2) of 0.79 and Nash-Sutcliffe Efficiency (NSE) criteria resulted in values of 0.68. The simulation which combines the hydrology model (HSPF) and the meteorological model (MM5) has shown better results than observation data. Model computation for coastal estuary modeling was also performed. The simulation provided a time series of sea water level, temperature, salinity, and distribution of sediment concentration in the Garang estuary, Semarang. The simulation result of sea surface elevation changes agree very well with the determination coefficient (R2) of 0.83 if compared with Semarang station, and 0.85 compare with Tanjungmas Port station. The result also showed that circulation in estuary has been dominated by river discharge rather than sea tide. For further research, it is suggested to consider the Journal of International Development and Cooperation, Vol.17, No.2, 2011, pp. 159-181

160

Adi PRASETYO, Takao YAMASHITA

interaction of meteorology-hydrology-coastal/estuary components by using the coupler system in the Earth system modeling (RES).

1. Introduction The phenomena of climate change is causing adverse impacts on human wellbeing and producing damage to the environment in many areas. IPCC (2007) argues that climate change is strongly affecting many aspects of the Earth system related to snow, ice and frozen ground (including permafrost); emerging evidence shows changes in the hydrological system, water resources, coastal zones and oceans. Climate change effects will likely increase water scarcity due to changes in the rainfall and hydrological pattern, increase the vulnerability of ecosystems due to temperature increase, frequent severe weather events and prolonged droughts and increase extreme precipitation and flooding, which will increase river water level, erosion rates and wash soil-based pollutants and toxins into waterways. Floods and inundations are considered as serious problems faced by coastal urban cities in the world. Indonesia, as an archipelago country, has a cumulatived coastal perimeter for more than twofold circumference of earth and is one of the countries vulnerable to the effect of climate change. As a tropical country located at the equator, Indonesia archipelago is recognized to have an important part in global circulation and climate system. Moreover, it is located between two continents, Asia and Australia, and two oceans the Pacific and Indian ocean. The dynamic interaction between atmospheric procesess and ocean in the Pacific ocean (ENSO) and the Indian ocean (DME) will influence surface temperature in Indonesian. Ashok et al. (2001) and Neale and Slingo (2003) stated that small changes in the sea surface temperature in maritime contingent such as Indonesia, can result significant changes in precipitation patterns across the Indo pacific region. During normal years, the variation of sea surface temperature (SST) and sea surface salinity (SSS) anomalies in the Java Sea are about + 0.5oC and + 0.2 psu, respectively. SST decreases (SSS increase) during DME and El Nino events because of transport of watermass with cooler SST (higher SSS) from eastern Java Sea, while during the La Nina events, the SST increase (SSS decrease) because of the transport of watermass with warmer SST (lower SSS) from the eastern Java Sea (Putri, 2005). Land surface interaction with atmospheric circulation in Indonesia is also effected by the geographical position of Indonesia which is located between Asia and Australia continent. Wyrtki (1961), Petit (1996) and Sadhotomo (1996) had analysed seasonal variation of environment, climate and ocean dynamics in the Java Sea. Furthermore, regional dynamic changes of the climate system in Indonesia become important factors influencing the Earth global system. No generally accepted definition of the “Earth System” exists (Haggag, 2009), and following notation was put forward by Schellnhuber (1998): the Earth system encompasses the natural environment, i.e the climate system or sometimes referred to as the ecosphere and the anthroposphere (Peixoto and Oort, 1992 in Haggag, 2009). To learn on interaction between the Earth and its components, it is needed to have simulation models that are able to observe the Earth environment processes and feedbacks between the components which possibly influence the properties of the whole system (Haggag, 2009). Yamashita et al. (2007) had developed the Asian Environment Simulator (AES) that was improved to a Regional Environment Simulator (RES) by Haggag (2009) to simulate mesoscale water, material transport and energy circulation in continental and coastal area. The Regional Environment Simulator (RES) is a coupled system of computer simulation for meteorology, physical oceanography, land surface, vegetation, hydrology, coastal dynamics and urban environment which is mainly used for natural environmental assessment against human activities (Yamashita et al, 2007). Lee (2006) had developed the meteorological model to provide climatology data fields for the atmosphere-ocean model couple. Kuroki (2007) and Wahyu (2009) had applied AES/RES in their research related to meteorological and hydrological variability in subtropical and tropical basin considering climate variation.

3


161

had applied AES/RES in their research related to meteorological and hydrological variability in subtropical and tropical basin considering climate variation. Marfai (2003; 2008) had integrated Marfai (2003; 2008) had integrated hydrologic in Semarang and some other researches related to GIS and hydrologic modelingGIS in and Semarang and modeling some other researches related to coastal dynamics coastal dynamics in the north coast of Java Island. in the north coast of Java Island. it is important to understand thecomponents interactingofcomponents of the Since itSince is important to understand the interacting the Earth system andEarth globalsystem climate and changes globalvariability climate and changes including and changes in Indonesia, andinthe limited including seasonal changes variability in Indonesia, andseasonal the limited researches into this topic coastal cities in researches into researches this topic which in coastal cities Indonesia, especially researches consider theEarth Indonesia, especially consider thein dynamic interactions among the majorwhich components of the dynamic interactions among the major components of the Earth system, as atmosphere, land system, as atmosphere, land surface/hydrology and the coastal/estuary, this study was carried out. surface/hydrology and the coastal/estuary, this study was carried out. The aim of this study is to evaluate and validate the applicability of several models as part of the Regional The aim of this study is to evaluate and validate the applicability of several models as part Environment Simulator (RES) on Atmospheric, Hydrological Coastal Circulation in Coastal Semarang of the Regional Environment Simulator (RES) onand Atmospheric, Hydrological andCity, Coastal Indonesia, such as non-hydrostatic meteorological model (MM5), watershed simulation model (HSPF), and Circulation in Coastal City, Semarang Indonesia, such as non-hydrostatic meteorological modelthreedimensional with sediment transport model forthree-dimensional estuary and coastal ocean (ECOMSED). (MM5),hydrodynamic watershed simulation model (HSPF), and hydrodynamic with sediment transport and coastal ocean (ECOMSED). By usingmodel RES asfor theestuary Earth system modeling, future climate variability and change; including the possible effects By using as climate the Earth system climate and change; of human activites on theRES global system, can bemodeling, predicted. Itfuture is expected thatvariability this study becomes one of the including the possible effects of human activites on the global climate system, can be predicted. It effective environmental assessment tools for regional integrated sustainable development plans in Indonesia, especially is expected that this study becomes one of the effective environmental assessment tools for regional in coastal cities. integrated sustainable development plans in Indonesia, especially in coastal cities. 2. Description of Study Area Area 2. Description of Study

Semarang is the capital city of Central Java Province and the fifth biggest city in Indonesia. o in Indonesia. Semarang is the Semarang capital citylies of Central the fifth biggest Geographically, on the Java northProvince coast atand coordinates EL 109city 35‘– 110o 50‘ Geographically, and SL 6o o o o o o Semarang on the coast at coordinates ELis109 35‘– 110 50‘ and SLrepresenting 6 50’ – 7 10’. Thenorth topography of Semarang divided into two parts the The city topography contour, of 50’ – lies 7 10’. namely the lowinto lying the north and highland the south. Thelying location Semarang is in Semarang is divided twoarea partsinrepresenting the city contour,innamely the low area inofthe north andarea highland shown in Figure 1. the south. The location of Semarang area is shown in Figure 1.

Figure 1: Map of Semarang

Figure 1: Map of Semarang As a tropical city, Semarang is situated in a tropical humid climate characterized by two seasons, dry and As a tropical city, Semarang is situated in a tropical humid climate characterized by two wet season. The climate is influenced by monsoon high precipitation. Several meteorological stations record the seasons, dry and wet season. The climate is influenced by monsoon high precipitation. Several precipitation data of Semarang area, such as the Simongan, Semarang, Gunungpati, Ungaran and Sumurjurang stations. meteorological stations record the precipitation data of Semarang area, such as the Simongan, Location of each meteorological station is depicted in Figure 3. Based on precipitation data from these stations in 1997-2006, except 1998,1999 and 2002, annual precipitation rate in Semarang was high, varying between 1,000 to 4,000 mm/year with an average of 2,655 mm/year for Simongan, 2,293 mm/year for Semarang, 2,554 mm/year for

Semarang, Gunungpati, Ungaran and Sumurjurang stations. Location of each meteorological station is depicted in Figure 3. BasedAdionPRASETYO, precipitation data from these stations in 1997-2006, except Takao YAMASHITA 162 1998,1999 and 2002, annual precipitation rate in Semarang was high, varying between 1,000 to 4,000 mm/year with an average of 2,655 mm/year for Simongan, 2,293 mm/year for Semarang, 2,554 mm/year for Gunungpati, 2,623 mm/year for Ungaran and 2,701 mm/year for Sumurjurang Gunungpati, 2,623 mm/year for Ungaran and 2,701 mm/year for Sumurjurang station. The annual precipitation rate for station. The annual precipitation rate for each meteorological station in Semarang city is shown in each meteorological station in Semarang city is shown in Figure 2 (a). Figure 2 (a).

(a)

(b)

Figure 2: (a) Annual precipitation and Figure 2: (a) Annual precipitationstation and in Semarang City (b) Average monthly precipitation for each meteorological

(b) Average monthly precipitation for each meteorological station in Semarang City Figure 2 (b) shows the total monthly precipitation from 1997, 2000, 2001, 2003, 2005 and 2006 varying between the and total monthly precipitation from 1997, 2001, 200 mm -Figure 500 mm2in(b) theshows wet season 0 mm – 50 mm in the dry season. This figure2000, also shows the2003, season2005 patternand in 2006 varying between 200 mm 500 mm in the wet season and 0 mm – 50 mm in the dry season. Semarang that is the rainy season from November-March and the dry season from June to August. This Semarang, figure also the season pattern frequently in Semarang that flood is the season from as oneshows of the coastal cities in Indonesia, suffers from whichrainy is typically caused by November-March and the dry season from June to August. upper course overflow of one river (river flood), drainage overflow of lowland area due to high rainfall intensity (local Semarang, of ‘rob’ the in coastal citiesBecause in Indonesia, from flood which is flood), and tidal floodas (orone named Indonesia). the Garangfrequently River as the suffers largest river taking an important typically caused by upper course overflow of one river (river flood), drainage overflow of lowland role in the development plan of Semarang city, this study will focus on the catchment of the Garang watershed. The area due to high rainfall intensity (local flood), and tidal flood (or named ‘rob’ in Indonesia). boundary area of the Garang watershed as shown in Figure 4 is made based on the landcover map from the Indonesia Because theForestry GarangandRiver as thebylargest river taking an important role in the development plan of Ministry of deliniated using Arc View GIS version 9.2 and Watershed Modeling System (WMS) Semarang this study will focus on the catchment the Garang boundary version 7.0.city, The Garang watershed width takes 211.3 km2 of theof catchment area orwatershed. about 56,4%The of Semarang areaarea as ofshown the inGarang watershed as shown in Figure 4 is made based on the landcover map from the Table 1.

Indonesia Ministry of Forestry and deliniated by using Arc View GIS version 9.2 and Watershed Modeling SystemTable (WMS) The of Garang watershed takes 211.3 km2 of the 1: Totalversion width and7.0. percentage landuse in the Garang width watershed in 2006 catchment area or about 56,4% of Semarang area as shown in Table 1. Landuse utilized

Area (km2)

Percentage (%)

Forest 25.64in the Garang watershed 12.13 Table 1: Total width and percentage of landuse in 2006 2 ) Percentage (%) Landuse utilized Area (km Plantation 2.33 1.1 Forest 25.64 Residential 50.76 24.02 12.13 Plantation 2.33 1.1 Dryland Farming 57.76 27.33 Residential 50.76 24.02 Mix Dryland Farming 66.82 31.62 Dryland Farming 57.76 27.33 Paddy Field 8.03 3.8 Mix Dryland Farming 66.82 31.62 Total 211.3 8.03 100 Paddy Field 3.8 Total 211.3 100

5 Numerical Study on Atmospheric, Hydrological and Coastal Circulation in Coastal City, Semarang Indonesia

Semarang Simongan

Gunungpati Sumurjurang Ungaran

Figure3:3:Location Location of stations in Semarang Figure of meteorological meteorological stations in Semarang

3

4 6 Garang river 2 3

5

1

4

River 1 2 3 4 5 6

Figure 4: Spatial distribution of the Garang watershed landuse in 2006

163

6

164


Figure 4: Spatial distribution of the Garang watershed landuse in 2006 3. Method 3. Method In this study, different models as part of the Earth system modeling, the Regional Environment Simulator (RES) were applied in order to achieve the objectives as given in the introduction. For atmospheric the Fifth Generation Mesoscale Meteorogical Model(RES) In this study, different modelsmodeling, as part of the Earth system modeling, the Regional Environment Simulator (MM5) was used while the Hydrological Program Fortran wasmodeling, applied the forFifth were applied in order to achieve the objectivesSimulation as given in the introduction. For (HSPF) atmospheric watershed simulation. For coast and estuary, hydrodynamic simulation used the ECOMSED Generation Mesoscale Meteorogical Model (MM5) was used while the Hydrological Simulation Program Fortran (Estuary Ocean Model with Sediment (HSPF) and was Coastal applied for watershed simulation. For coastTransport). and estuary, hydrodynamic simulation used the ECOMSED (Estuary and Coastal Ocean Model with Sediment Transport).

3.1 Model description MM5 3.1 Model description MM5 is one of the meteorological/atmospheric models designed to simulate mesoscale and MM5 regional scale weather as closely as possible with actual atmospheric condition. MM5 was MM5 is one of the meteorological/atmospheric models designed to simulate mesoscale and regional scale developed in 1970 by Pennsylvania State University and National Center for Atmospheric Research weather as closely as possible with actual atmospheric condition. MM5 was developed in 1970 by Pennsylvania State (PSU/NCAR) of U.S. This model is one of the most widely used three-dimensional prognostic University and National Center for Atmospheric Research (PSU/NCAR) of U.S. This model is one of the most widely meteorological models. used three-dimensional prognostic meteorological models. According to Grell et al. (1996), MM5 characteristics are functional in a According to Grelland et al. non-hydrostatic (1996), MM5 characteristics functionaldesigned in a regional/limited-area regional/limited-area model are primarily to simulateandornon-hydrostatic predict model primarily designed to simulate or predict mesoscale/regional atmospheric circulation. The model mesoscale/regional atmospheric circulation. The model is supported by several pre-is supported and by several pre- and post-processing which referred to collectively as the MM5 modeling and also post-processing programs, whichprograms, are referred toare collectively as the MM5 modeling systemsystem and also uses a terrain following non dimensional pressure or sigma coordinates pressure. Schematic MM5 modeling is shown uses a terrain following non dimensional pressure or sigma coordinates pressure. Schematic MM5 in Figure is 5. shown in Figure 5. modeling

Figure5:5:MM5 MM5 modeling modeling system flowchart Figure system flowchart

7 7

Numerical Study on Atmospheric, Hydrological and Coastal Circulation in Coastal City, Semarang Indonesia 7

7

7 7

165

The equationofof MM5 The governing governing equation MM5 According toGrell Grell (1996), vertical σ-coordinate is defined in of terms of pressure as: The governing equation of MM5 σ-coordinate According to etetal.al. (1996), vertical is defined in terms pressure as: The governing equation MM5vertical σ-coordinate is defined in terms of pressure as: According to Grell et al. of (1996), The governing equation of MM5vertical σ-coordinate is defined in terms of pressure as: According to Grell al.of (1996), The governing equation MM5 et ...(1) According to Grell et al. (1996), vertical σ-coordinate σ-coordinate is is defined defined in in terms terms of of pressure pressure as: as: The governing equation of MM5 According to Grell et al. (1996), vertical According to Grell et al. (1996), vertical σ-coordinate is defined in terms of pressure as: ...(1) where statepressure, pressure,ptpat aspesified spesified constant pressure, ps reference the reference constant toptop pressure, ps the state state surface pressure where p p is reference reference state ...(1) surface the levels in non-hydrostatic system fixed in space closely ...(1) and theppressure σislevels inand non-hydrostatic system fixed in space are closely related height ratherrelated than constant top pressure, ptos are the reference statepressure and where reference stateσ pressure, pt a spesified ...(1) to than a certain point, isspace calculated by p related = ...(1) σ(ps of pressure where ppressure israther reference state pnon-hydrostatic a spesified constant top pressure, pp's is thea closely reference state t pressure surface and pressure the σpressure, levels inthe system in are theheight pressure at a certain point, p,and is calculated by p at = σ(p pfixed p'p, where computed value s - pt) + t + where p is reference state pressure, p constant top pressure, p state t a spesified s the reference ...(1) -where p ) + p + p' where p' is a computed value of pressure perturbation. surface and state the σpressure, levels systemtop fixed in is space are reference closely theight ptpressure israther reference pnon-hydrostatic spesified pressure, ps the ta to than pressure andinthe pressure at constant a certain point, p, calculated by prelated = state σ(ps perturbation. surface andstate the σ levels levels inpthe non-hydrostatic system fixed in is space arereference closely related where p ispressure reference avalue spesified toppoint, pressure, ps the tnon-hydrostatic to rather than pressure andin pressure atconstant a certain p, calculated by prelated =state σ(ps surface and system fixed in space are closely - pheight p' where p'the is σ apressure, computed of pressure perturbation. t) + ppressure t+ to height rather thanthe pressure and the pressure at aasystem certain point, p,and is calculated calculated by prelated psystem = σ(p σ(ps surface and in the non-hydrostatic fixed space closely The equations inofCartesian-sigma (x, in yp, σ)are coordinate - pheight + pressure pt +rather p' where p'governing isσalevels computed value pressure perturbation. to than pressure and pressure at certain point, is by = t) non-hydrostatic s The non-hydrostatic governing equations in Cartesian-sigma (x, y and σ) coordinate system in a rotating frame of ) + p + p' where p' is a computed value of pressure perturbation. p toin thanofpressure andcan thebepressure at certain point, calculated by p =system σ(ps reference written asafollows: + pttrather + p'frame where p'governing is a computed value pressure perturbation. - height patt)rotating The non-hydrostatic equations inofCartesian-sigma (x,p,y isand σ) coordinate reference can be written as follows: - Pressure pt)a+rotating pt + p' where isreference a computed value of perturbation. The non-hydrostatic equations in pressure Cartesian-sigma (x, y and σ) coordinate system in frame p' ofgoverning can be written as follows: The non-hydrostatic governing equations in Cartesian-sigma Cartesian-sigma (x, yy and and σ) σ) coordinate coordinate system system Pressure in a rotating frame of reference can be written as follows: The non-hydrostatic governing equations in (x, Pressure in a rotating frame of reference can be written as follows: The governing in Cartesian-sigma Pressure in anon-hydrostatic rotating frame of referenceequations can be written as follows: (x, y and σ) coordinate system Pressure ...(2) ...(2) inPressure a rotating frame of reference can be written as follows: Momentum (x-component) Pressure ...(2) ...(2) Momentum (x-component) Momentum (x-component) ...(2) Momentum (x-component) ...(2) Momentum (x-component) ...(3) ...(2) Momentum (x-component) ...(3) Momentum(x-component) (y-component) Momentum ...(3) ...(3) Momentum (y-component) ...(3) Momentum (y-component) ...(3) Momentum (y-component) Momentum (y-component) (y-component) ...(4) ...(3) Momentum Momentum(y-component) (z-component) Momentum ...(4) ...(4) ...(4) Momentum (z-component) ...(4) Momentum (z-component) ...(4) Momentum (z-component) Momentum (z-component) ...(5) ...(4) Momentum (z-component) Thermodynamics Momentum (z-component) ...(5) ...(5) ...(5) Thermodynamics ...(5) Thermodynamics ...(5) Thermodynamics ...(5) Thermodynamics ...(6) Thermodynamics Thermodynamics where T, ρ, θ, Q and D( ) are the temperature, density, potential temperature, heating...(6) rate due to diabatic eddy density, viscous terms respectively. The heating constants g, ...(6) ...(6) where T, ρ, θ, processes Q and D( and ) aresubgrid-scale the temperature, potential temperature, rate ...(6) f,due Rd,tocp areand theD( acceleration due to eddy gravity, Coriolis parameter, the gas constant ofg, where T, and ρ, θ,γ Q ) aresubgrid-scale the temperature, density, potential temperature, rate ...(6) diabatic processes and viscous terms respectively. Theheating constants where T, ρ,heat θ, Q Qcapacity and D( D( of are the temperature, temperature, density, potential temperature, heating dry air, theand constant pressure of density, air Coriolis and the ratio of heatthe capacity for...(6) airrate at due to cp diabatic subgrid-scale viscous terms respectively. The constants g,to diabatic where T, ρ, θ, and ))are are the potential temperature, f, Rd, γQprocesses are the due to eddy gravity, parameter, gas heating constant of where T, ρ , θ, and D( acceleration )and the temperature, density, potential temperature, heating rate rate due due to diabatic processes and subgrid-scale eddy viscous terms respectively. The constants g, where T, ρ, θ, Q and D( ) are the temperature, density, potential temperature, heating rate constant pressure to the thatacceleration of constant volume. Here, the two new parameters, α, which isair the f, Rd, and γprocesses are due to eddy gravity, Coriolis parameter, the gas constant of due to cp diabatic and subgrid-scale viscous terms respectively. The constants g, dry air, the heat capacity of constant pressure of air and the ratio of heat capacity for at processes and subgrid-scale eddy viscous terms respectively. The constants g,and f, the Rd, cp and γ are the acceleration f, Rd, cp and γ are the acceleration due to gravity, Coriolis parameter, gas constant of due to diabatic processes and subgrid-scale eddy viscous terms respectively. The constants g, angular difference between the y-axis of the grid and the true north e = 2Ωcosφ (note dryRd, air,cpthe heat capacity ofconstant constantdue pressure of airthe andtwo thenew ratio of heatthe capacity for is airthe at f, and γ are to gravity, Coriolis parameter, gas constant of constant pressure tothe thatacceleration of volume. Here, parameters, α, which dry air, the heat capacity of constant pressure of air and the ratio of heat capacity for air at due tothe parameter, gas constant of dry air, the heat capacity of capacity constant pressure of air and the f,with Rd,air, cpgravity, and γCoriolis arecapacity the acceleration due gravity, Coriolis parameter, the constant of internal energy, e), arethe introduced toHere, describe the full three-dimensional Coriolis constant pressure to that of volume. the two new parameters, which is the dry the heat of constant constant pressure of air and the ratio of heat for air at angular difference between the y-axis oftothe grid and the true north and egas =α,2Ωcosφ (note constant pressure to that of constant volume. Here, the two new parameters, α, which is the ratio of heat capacity for air at constant pressure to that of constant volume. Here, the two new parameters, α, dry air, the heat capacity of constant pressure of air and the ratio of heat capacity for air at effects where Ω and φ are the angular velocity of the Earth and the latitude. angular the of thetoHere, grid and and e =α,2Ωcosφ (note constant pressure tobetween that of volume. the the two new parameters, which is the with thedifference internal energy, e),constant arey-axis introduced describe thetrue fullnorth three-dimensional Coriolis angular difference between the y-axis of the grid and the true north and e = 2Ωcosφ (note constant pressure to that of constant volume. Here, the two new parameters, α, which is the which is the angular difference between the y-axis of the grid and the true north and e = 2Ωcosϕ (note with the with thewhere internal energy, are introduced fullthe three-dimensional Coriolis angular difference between y-axis of thetogrid andEarth thethetrue north and e = 2Ωcosφ (note effects Ω and φ are e), thethe angular velocity ofdescribe the and latitude. with the internal energy, e), to full Coriolis angular difference between theare y-axis the and thethe true north and e effects = 2Ωcosφ (note Advection terms can expressed as of internal energy, areφbe introduced tointroduced describe thegrid three-dimensional Coriolis where Ω and ϕ are the with the internal energy, e), are introduced tofull describe the full three-dimensional Coriolis effects where Ωe),and are the angular velocity ofdescribe the Earth and thethree-dimensional latitude. effects where Ω and and φ are aree),the the angular velocity of the the Earth Earth and the latitude. with the internal energy, are introduced to describe the full three-dimensional Coriolis effects where Ω φ angular velocity of and the latitude. angular velocity the Earth and the latitude. Advection termsofcan be expressed as effects whereterms Ω and φ be areexpressed the angular Advection can asvelocity of the Earth and the latitude. Advection terms terms can can be be expressed expressed as as ...(7) Advection Advection terms can velocity, be expressed as where the vertical , in σ coordinate is related to modeled values of velocity ...(7) Advection terms can be expressed as ...(7) where the vertical velocity, , in σ coordinate is related to modeled values of velocity ...(7) ...(7) where the vertical velocity, , in σ coordinate is related to modeled values of velocity ...(7) where the vertical velocity, , in σ coordinate is related to modeled values of velocity ...(7) where the vertical velocity, , in σ coordinate is related to modeled values of velocity where the vertical velocity, , in σ coordinate is related to modeled values of velocity

8

166


8

・

where the vertical components by velocity, σ , in σ coordinate is related to modeled values of velocity components by components by

...(8)

...(8) ...(8)

Divergence can be beexpanded expandedinin Cartesian-Sigma coordinate Divergence term term can thethe Cartesian-Sigma coordinate as as Divergence term can be expanded in the Cartesian-Sigma coordinate as

...(9) ...(9) when p* is constant, the second and fourth terms on right hand side will disappear....(9) On when p* constant, the second fourth onhand rightside hand side will disappear. On surface (σ=0) and topthe(σ=1) levels, boundaries considered as rigid conditions so that when p* isisconstant, second and and fourth terms terms onare right will disappear. On surface (σ =0) and top ・ (σ=0) and top (σ=1) levels, boundaries are considered as rigid conditions so that =surface 0 is applied. (σ=1) levels, boundaries are considered as rigid conditions so that σ = 0 is applied. = 0 is applied. In order to improve the rainfall simulation results, Four Dimensional Data Assimilation In order totoimprove the rainfall simulation results,results, Four Dimensional Data Assimilation (FDDA) was employed order improve rainfall simulation Dimensional Data to Assimilation (FDDA) In was employed in thetheMM5 simulation. FDDA is aFour method and technique assure a in the MM5 simulation. FDDA a method and technique to is assure a dynamical consistencytothat keep the (FDDA) was employed in keep the isMM5 simulation. FDDA a method and technique dynamical consistency that the model as closely to actual atmoshperic conditions assure and is a model as closely to actual atmoshperic conditions and is making up for erors and gaps in the initial analysis and deficiencies in dynamical consistency that keep the model as closely to actual atmoshperic conditions and is making up for erors and gaps in the initial analysis and deficiencies in the physics model. the physics model. making up for erors and gaps in the initial analysis and deficiencies in the physics model. HSPF HSPF Hydrological Simulation Program Fortran (HSPF) is a one-dimensional comprehensive HSPF Hydrological Simulation Program (HSPF) is a one-dimensional comprehensive watershed simulation Simulation model developed byFortran theFortran United States Protection Agency (US simulation Hydrological Program (HSPF) is a Environmental one-dimensional comprehensive watershed watershed simulation modelfrom developed by the United States Environmental Protection Agency (US EPA). This model is derived the Standford Watershed Model (SWM) and designed to simulate model developed by the United States Environmental Protection Agency (US EPA). This model is derived from the This modeland is derived from the Standford Watershed Modeland (SWM) and designed tosediment simulate allEPA). water quantity quality processes that occur in a watershed in-stream including Watershed Model (SWM) and designed to simulate all water quantity and quality processes that occur in allStandford waterand quantity and quality processes that occur in reproduce a watershed and in-stream including sediment transport movement of contaminants. HSPF can spatial variability by dividing the a watershed in-streamofincluding sediment transport and movement of contaminants. by HSPF can reproduce spatial transport and and movement contaminants. HSPF can the basin in hydrological homogenous land segments andreproduce simulatingspatial runoffvariability for each landdividing segment, variability by dividing the basin in hydrological homogenous land segments and simulating runoff for each land basin in hydrological homogenous land segments and and simulating runoff for each(Bicknell land segment, independently, using different meteorological input data watershed parameters et al. segment, independently, using different meteorological input data and watershed parameters (Bicknell et al. 1996). independently, using different meteorological input data and watershed parameters (Bicknell et al. 1996). HSPF includes parameterizations for the main hydrological process (e.g. precipitation, 1996). includes parameterizations for the main hydrological process (e.g. precipitation, HSPF HSPF includes parameterizations for the main hydrological process (e.g. precipitation, interception, evapotranspiration, interception, evapotranspiration, surface runoff, ground water, interflow, etc). interception, evapotranspiration, surface runoff, ground water, interflow, etc). surface runoff, water,stages interflow, There are ground three main thatetc). should be done in order to run this simulation model: preThere arethree threeand main stages should be done in toorder to run this simulation model: pre- processing processing, processing post processing. There are main stages thatthat should be done in order run this simulation model: preprocessing, processing, processing and post processing. 1. In the pre-processing step, all data on topography, weather (precipitation, temperature, and post processing. 1. etc) In the pre-processing all dataofonstudy topography, (precipitation, landuse in step, thestep, watershed area areweather required. Landuse on temperature, this 1. In theand pre-processing all data on topography, weather (precipitation, temperature, etc)site andislanduse in the etc) and landuse in the watershed of study area are required. Landuse on this analysed by ArcGIS software version 9.2. ArcGIS is a tool to prepare the land use site mapis watershed ofbystudy area are required. Landuse onArcGIS this site is is analysed by ArcGISthe software version analysed ArcGIS software version 9.2. a tool to prepare land use map9.2. ArcGIS and also to convert 90-meter resolution Digital Elevation Model (DEM) from isand a toolalso to prepare the land use map and also to convert 90-meter resolution Digital Elevation convert 90-meter resolution Digital Elevation Model derivatives (DEM) from Hydrosheds to(Hydrological data and maps based on shuttle elevation atModel (DEM) from Hydrosheds (Hydrological data covering and based on shuttle derivatives at multiple Hydrosheds data andmaps maps based on elevation derivatives multiple scale).(Hydrological The landuse map the whole areashuttle ofelevation Semarang, specifically theat scale). The landuse map covering the wholemap area of Indonesia Semarang, specifically theSemarang, Garang was the obtained from multiple scale). The landuse covering the whole area Garang watershed, was obtained from Ministry of of Forestry in watershed, thespecifically year 2006. Garang watershed, was from2006. Indonesia Ministry of Forestryparameters in the yearhave 2006.to Indonesia Ministrystage of Forestry in the year 2. The processing usesobtained the WinHSPF model. Many hydrological 2. processing uses the hydrological parameters to in stage this stage. TheWinHSPF model hasmodel. three Many main modules: PERLND, 2. be TThe heconsidered processing stage uses the WinHSPF model. Many hydrological parameters have IMPLND, to behave considered in this be considered in this stage. The model has three main modules: PERLND, IMPLND, and RCHRES which are representing pervious land segments, impervious land stage. The model has three main modules: PERLND, IMPLND, and RCHRES which are representing pervious and RCHRES which reaches/mixed are representing pervious land segments, impervious land segments, andimpervious free-flow reservoirs, land segments, land segments, and free-flowrespectively. reaches/mixed reservoirs, respectively. segments, and free-flow reaches/mixed reservoirs, respectively. 3.3. The post processing is the last process in hydrological modeling. Output result or The post processing is the last process in hydrological modeling. Output result or simulation can be produced 3. simulation The post processing is the last process in hydrological result or can be produced by using the hydrograph. Inmodeling. this study,Output the Scenario bysimulation using the hydrograph. In this study, the Scenario Generator (GenScn) was used. the The Scenario GenScn provides the can be produced by using the hydrograph. In this study, Generator (GenScn) was used. The GenScn provides the ability to change an input ability to change an inputwas sequence interactively, to run the HSPF model, and view results graphically. The Generator (GenScn) used. The GenScn provides ability toto change an The input sequence interactively, to run the HSPF model, and to the view results graphically. results of different scenarios can be easily compared and analyzed because the model and analysis sequence interactively, to run the HSPF model, and to view results graphically. The tools are results of different scenarios can be easily compared and analyzed because the model results of different scenarios can be easily compared and analyzed because the model linked in one package and use a common database. and analysis tools are linked in one package and use a common database. and analysis tools are linked in one package and use a common database.

9 99 9 Numerical Study on Atmospheric, Hydrological and Coastal Circulation in Coastal City, Semarang Indonesia

167

ECOMSED ECOMSED ECOMSED ECOMSED

ECOMSED ECOMSED (Estuary and Coastal Ocean Model with Sediment Transport) the ECOMSED (Estuary and Coastal Ocean Model with Sediment Transport) isone one the ECOMSED (Estuary and Coastal Ocean Model with Sediment Transport) isone one ofof the ECOMSED (Estuary and Coastal Ocean Model with Sediment Transport) isis ofof the ECOMSED (Estuary and Coastaldesigned Ocean Model with Sediment Transport) ishydrodynamic one of the comprehensive comprehensive coastal-estuary models to simulate three-dimensional in the comprehensive coastal-estuary models designed to simulate three-dimensional hydrodynamic in comprehensive coastal-estuary models designed to simulate three-dimensional hydrodynamic in the comprehensive coastal-estuary models designed to simulate three-dimensional hydrodynamic in thethe coastalmarine and fresh water system. This model from the Princeton Ocean Model (POM) estuary models designed to simulate three-dimensional hydrodynamic in the marine and fresh water system. This model marine and fresh water system. This model isdeveloped developed from the Princeton Ocean Model (POM) marine and fresh water system. This model isdeveloped developed from the Princeton Ocean Model (POM) marine and fresh water system. This model isis from the Princeton Ocean Model (POM) since 1980. The complete ECOMSED model consists of several modules, these are hydrodynamic since 1980. The complete ECOMSED model consists of several modules, these are hydrodynamic since 1980. The complete ECOMSED model consists of several modules, these are hydrodynamic is developed from the Princeton Ocean Model (POM) since 1980. The complete ECOMSED model consists of several since 1980. The complete ECOMSED model consists of several modules, these are hydrodynamic module, sediment transport module, wind induced wave module, heat flux module and particle module, sediment transport module, wind induced wave module, heat flux module and particle module, sediment transport module, wind induced wave module, heat flux module and particle modules, these are hydrodynamic module, sediment transport module, wind induced wave module, heat flux module module, sediment transport module, wind induced wave module, heat flux module and particle tracking module (Blumberg and Mellor, 1987). The modules within the ECOMSED modeling tracking module (Blumberg and Mellor, 1987). The modules within the ECOMSED modeling tracking module (Blumberg and Mellor, 1987). The modules within the ECOMSED modeling tracking module (Blumberg and Mellor, 1987). The modules within the ECOMSED modeling and particle tracking module (Blumberg and Mellor, 1987). The modules within the ECOMSED modeling framework framework are linked internally. These modules be turned on and off by the users depending framework are linked internally. These modulescan can turned on and the users depending framework are linked internally. These modules can bebe turned onusers and offoff byby the users depending framework are linked internally. These be turned on and off by the users depending are linked internally. These modules canmodules be turnedcan on and off by the depending upon their needs. The ECOMSED upon their needs. The ECOMSED modeling framework also allows for linking each module upon their needs. The ECOMSED modeling framework also allows for linking each module upontheir theirneeds. needs.The TheECOMSED ECOMSEDmodeling modelingframework frameworkalso alsoallows allowsfor forlinking linkingeach eachmodule module upon modeling framework also allows for linking each module externally. externally. externally. externally. externally. ECOMSED potentialto totoprovide provide a time series of water circulation, temperature, salinity,salinity, number and diameter ECOMSED potential series water circulation, temperature, ECOMSED isis potential provide atime time series water circulation, temperature, salinity, ECOMSED ispotential potential provide time series ofof water circulation, temperature, salinity, ECOMSED isis totoprovide aaatime series ofof water circulation, temperature, salinity, of particle, and deposition and re-suspension of cohesive and non-cohesive sediments as non-cohesive closely as possible with the number and diameter and deposition and re-suspension and number and diameter ofparticle, particle, and deposition and re-suspension ofcohesive cohesive and non-cohesive number and diameter ofparticle, particle, and deposition and re-suspension ofcohesive cohesive and non-cohesive number and diameter ofof and deposition and re-suspension ofof and non-cohesive sediments as closely as possible with the observation data. observation data. sediments as closely as possible with the observation data. sediments as closely as possible with the observation data. sediments as closely as possible with the observation data. The governing equation ECOMSED The governing equation of ECOMSED The governing equation ECOMSED The governing equation ofof ECOMSED The governing equation ofof ECOMSED According to Bryan (1969), the continuity equation is: According to Bryan (1969), the continuity equation According to Bryan (1969), the continuity equation is:is: According to Bryan (1969), continuity equation is: According to Bryan (1969), thethe continuity equation is:

...(10)

...(10) ...(10) ...(10) ...(10) where∇∇isis the the horizontal velocity vector withwith components (U, V)(U, andV) ∇and the horizontal gradient operator and z horizontal velocity vector components ∇ the horizontal where where ∇ is the horizontal velocity vector with components (U, V) and ∇ the horizontal is the horizontal velocity vector with components (U, V) and ∇ the horizontal where where ∇ is the upwards. horizontal velocity vector with components (U, V) and ∇ the horizontal increasing vertically gradient operator and increasing vertically upwards. gradient operator and z increasing vertically upwards. gradient operator and increasing vertically upwards. gradient operator and zzzincreasing vertically upwards. The Reynolds momentumequations equations are: The Reynolds momentum are: The Reynolds momentum equations are: The Reynolds momentum equations are: The Reynolds momentum equations are:

...(11)

...(11) ...(11) ...(11) ...(11) with ρo the reference density, ρ in situ density, g gravitational acceleration, P the pressure, with ρo the reference density, ρinsitu insitu situ density, ggravitational gravitational acceleration, Pthe the pressure, with ρo the reference density, ρ density, g acceleration, P pressure, with ρo the reference density, ρ in density, g gravitational acceleration, P the pressure, with ρo the reference density, ρ in situ density, g gravitational acceleration, P the pressure, and KM the vertical and K the vertical eddy diffusivity momentum mixing. latitudinal variation the vertical eddy diffusivity ofturbulent turbulent momentum mixing.A latitudinal variation ofthe the and Kthe the vertical eddy diffusivity ofturbulent turbulent momentum mixing. AA latitudinal variation ofthe the and KMM M M and Keddy vertical eddy diffusivity ofof momentum mixing. variation ofof diffusivity of turbulent momentum mixing. Aplane latitudinal variationAoflatitudinal the Coriolis parameter, f,can iscan Coriolis parameter, f, is introduced by using the β approximation. The pressure at depth z Coriolis parameter, f, is introduced by using the β plane approximation. The pressure at depth z Coriolisparameter, parameter,f,f,isisintroduced introducedby byusing usingthe theββplane planeapproximation. approximation.The Thepressure pressureatatdepth depthz zcan canintroduced by Coriolis using the by β plane approximation. The pressure at depth zthe can be obtained by integrating the vertical component of the be obtained integrating the vertical component equation motion, from free obtained integrating the vertical component the equation motion, from tothe the free bebe obtained byby integrating the vertical component ofof the equation ofof motion, from tothe the free be obtained by integrating the vertical component ofof the equation ofof motion, from zzztozto free equation of motion, from z to the free surface, η, and is: surface, η, and is: surface, η, and is: surface, η, and is: surface, η, and is:

...(12)

...(12) ...(12) ...(12) ...(12) Henceforth, the atmospheric pressure P is assumed to be constant. The conservation is assumed to be constant. The conservation equations Henceforth, atmospheric pressure P atm Henceforth, the atmospheric pressure P is assumed to be constant. The conservation atm Henceforth, the atmospheric pressure P is assumed to be constant. The conservation atm atm Henceforth, the atmospheric pressure Patm is assumed to be constant. The conservationfor temperature 10 equations for temperature (θ): equations temperature (θ): equations forfor temperature (θ): (θ): equations for temperature (θ):

...(13)

...(13) ...(13) ...(13) ...(13)

andand forfor salinity salinity(S): (S):

...(14) where θ is the potential temperature (or in situ temperature for shallow water applications) and S is the salinity. The vertical eddy diffusivity for turbulent mixing of heat and salt is denoted as KH. Using the temperature and salinity, the density (ρ) is computed according to an equation of state of the form:

...(14)

...(14) where θ is the potential temperature (or in situ temperature for shallow water applications and for salinity (S): and S is the salinity. The vertical eddy diffusivity for turbulent mixing of heat and salt is denoted a Adi PRASETYO, Takao YAMASHITA 168 KH. Using the temperature and salinity, the density (ρ) is computed according to an equation o state of the form: ...(14)

where θ is the potential temperature (or in situ temperature for shallow water applications) where θ is the potential temperature (or in situ temperature for shallow water applications) and S is the salinity. and S is the salinity. The vertical eddy diffusivity for turbulent mixing of heat and salt is denoted as The vertical eddy diffusivity turbulentthe mixing of heat salt is denoted as KH. Using temperature KH. Using the temperature andforsalinity, density (ρ) and is computed according to anthe equation of and salinity, (ρ) is computed according to an equation of state of the form: statethe of density the form: 3.2 Validation and evaluation

...(15

some calibration and validation was done on all simulation models ...(15) In this study, in order t ...(15) adjust the simulation results as closely as possible with the observation data. For atmospheri 3.2 Validationvalidation and evaluation 3.2modeling, Validation and evaluation of the model was performed by comparing daily precipitation from In this study, some calibration validation wasMM5 done on all simulation modelsdata in order to meteorological stations in and Semarang with results. Re-analysis sattelite Tropica this study, some calibration and validation waswith done the on all simulation models in order tofrom adjust the the simulation adjust theInsimulation results as closely as possible observation data. For atmospheric Rainfall Measuring Mission (TRMM) was used as a comparison for daily rainfall results as closely asofpossible with thewas observation data.also atmospheric modeling, validation of the modeldistribution was modeling, validation the model performed byForcomparing daily precipitation from performedOn by comparing daily precipitation fromresults. meteorological stations Semarang with MM5 results. Re-analysis the performing ofwith theMM5 hydrological model, graphical and statistical methods were used t meteorological stations in Semarang Re-analysis datainfrom the sattelite Tropical data from the the sattelite Tropical Rainfall Measuring (TRMM) was as aby comparison for daily Rainfall Measuring Mission (TRMM) was also usedMission as a comparison for also dailyused rainfall distribution. evaluate model results. The hydrological model was calibrated comparing therainfall outflow of th distribution. On the performing of themodel hydrological and statisticaldata. methods were used to used three mode Garang River from resultsmodel, with graphical daily measurement This simulation Onmodel the performing of the hydrological model, graphical and of statistical methods were used toof evaluate the model evaluate the results. The model was calibrated byHSPF comparing the outflow the is the evaluation statistics tohydrological evaluate the performance model output, that coefficient o 2 simulation Garang River from model results with daily measurement data. This used three model results. The hydrological model was calibrated by comparing the outflow of the Garang River from model results correlation (R), the coefficient of determination (R ) which is defined as the squared with value of th evaluation statistics todata. evaluate the performance of model HSPF evaluation model output, thattoisevaluate the coefficient of daily measurement This simulation used three statistics theaperformance of HSPF coefficient of correlation, and the Nash-Sutcliffe Efficiency (NSE), normalized statistic tha 2 correlation (R), the coefficient of determination (R(R), ) which is definedofasdetermination the squared(R value of the 2 model output, that is the coefficient of correlation the coefficient ) which is defined as the determines the relative magnitude of theEfficiency residual (NSE), variance compared statistic to the measured data variance coefficient of correlation, and the Nash-Sutcliffe a normalized that squared value of the coefficient of correlation, and the Nash-Sutcliffe Efficiency (NSE), a normalized statistic that determines the relative magnitude of the residual variance compared to the measured data variance. Concerning validation of the coastal model, model results were validated by comparin determines the relative magnitude of the residual variance compared to the measured data variance. Concerning validation the measurements coastal model, model results were validated byarea. comparing calculated tidal levels of with of water the study A statistical analysis wa Concerning validation of the coastal model, model resultslevel were in validated by 2comparing calculated tidal calculated tidal levels with measurements of water level in the study area. A statistical analysis was perfomed as correlation of coefficient (R ). Aascomparison levels with measurements water (R) level dan in the determination study area. A statistical was perfomed correlation (R) of and sedimen perfomed as correlation (R)ofdan determination of coefficient (R2). analysis A comparison of sediment 2 was also verified by observed data. Due to limited observation dat concentration results determination of coefficient (R ).verified A comparison of sediment concentration also verifieddata by observed data. concentration results was also by observed data. Due to results limitedwas observation availability, some simulation outputs such as temperature and salinity distribution are validated b Due to limited observation data availability, some simulation outputs such as temperature and salinity are availability, some simulation outputs such as temperature and salinity distribution are validated bydistribution using by re-analysis data.data. using usingvalidated re-analysis data.re-analysis equation of evaluation criteriaisisshown shown below: The The equation of evaluation criteria below: The equation of evaluation criteria is shown below:

⎛ ⎞ n _ _ ⎛ n⎞⎛ ⎛⎛ ⎞_ ⎞⎛ ⎟ _ ⎜ ⎜ ⎛ ⎞ − − O O P P ⎜ ⎜ ⎟ ⎜ ⎟ ∑ i i ⎜ i =1 ⎝ ⎜ Oi − ⎠O ⎟⎜ ⎟Pi − P ⎟2 ⎜ ⎠ ⎝ ⎜ R =⎜ R⎠ = ⎜ ⎠⎝ ⎟ i =1 ⎝ _ 2 ⎟ ⎜ Rn ⎛= ⎜ _ ⎞ 2 ⎛ ⎜ ⎞2 _ 2⎜ ⎜ ∑ ⎜ O⎜i − On⎟ ⎛ ⎜ Pi −_P⎞⎟ ⎟ ⎛ ⎞ ⎠ ⎜ O⎝ − O ⎟⎠ ⎠ ⎜ P − P ⎟ ⎝ ⎝ i =1 ⎝ ⎜ i i

∑

n

NSE = 1 −

⎝

∑ (O i =1 n

i

∑⎝

⎠

i =1

− Pi )

2

n

∑ (O

_ ⎛ ⎞ ∑ NSE =⎜⎝ O1i −− Oi =⎟⎠1 i =1 n

2

i

− Pi )

2

_ ⎛ ⎞ O O − ⎜ ⎟ ∑ i ⎠ i =1 ⎝

2

⎝

2

⎞ _ _ ⎞ n ⎛ _ ⎟ ∑ ⎜ Oi − O ⎞⎟⎛⎜⎛ Pi − Pn ⎞⎟⎛ ⎟ _ ⎞⎛ ⎞ ⎠⎜⎝ ∑⎠⎜ Oi ⎟− O ⎟⎜ Pi − P ⎟ ⎟ i =1 ⎝ ⎟ ⎠ ⎟ n ⎛ R 2_ =⎞ 2⎜ ⎛ i =1 _⎝ ⎞ 2 ⎟ ⎠⎝ 2 _ 2 ⎟ ∑ ⎜ Oi − O ⎟ ⎜ ⎜ Pin −⎛P ⎟ ⎟ _ ⎞ ⎛ ⎞ ⎝ ⎠ ⎝ ⎠ i 1 = ⎟ ⎜ ∑ ⎜ Oi −⎠ O ⎟ ⎜ Pi − P ⎟ ⎠ ⎠ ⎠ ⎝ ⎠ ⎝ i =1 ⎝

...(16) ...(17)

⎞ ⎟ ⎟ ⎟ ⎟ ⎟ ⎠

2

...(16) ...(17)

...(16

...(17

where Oi and Pi are the observed and simulated data, and O and P are the mean observed and simulated during period. where data Oi and Pi arethe theevaluation observed and simulated data, and O and P are the mean observed and simulated data during thewhere evaluation O period. and P

i i are the observed and simulated data, and O and P are the mean observed an simulated data2 during the evaluation period.

The range of R and R between 0 and 1 will describe how much of the observed dispersion is explained by the prediction. A value of zero means no correlation at all, whereas a value of 1 means that the dispersion of the prediction is equal to that of the observation. NSE indicates how well the plot of observed versus simulated data fits the 1:1 line.


169

NSE ranges between –∞ and 1 with NSE = 1 being the optimal value. Value between 0 and 1 is generally viewed as acceptable level of performance (Moriasi et al., 2007 in Wahyu, 2009)

4. Results and Discussions 4.1 Atmospheric Modeling In this study, atmospheric simulation in Semarang area with several physics configuration had been performed with other scheme option using non and with observation nudging technique (FDDA). However, only the optimum simulation results, which were acquired from the modeling process, are presented in this paper. Detailed physics parameters for optimum combination option in MM5 simulation are shown in Table 2.

Table 2: Configuration of MM5 simulation MM5 Configuration

MM5

Simulation time Dynamics Input data

MM5+FDDA

January 1 – 30, 2005 Non Hdydrostatic NCEP FNL (1ox1o resolution)

Input Terrain data

USGS Global data

Number of domains

2 domains

Number of grids

Domain 1 : 35x41 Domain 2 : 49x52

Horizontal resolution of domain

Domain 1 : 27 km Domain 2 : 9 km

Atmospheric radiation scheme (FRAD) Planetary boundary layer parameterization (IBLTYP)

Cloud radiation (Dudhia scheme) Medium Range Forecast (MRF) Scheme

Land surface model (ISOIL)

NOAH

Cumulus parameterization (ICUPA)

Grell

Kain-Fritz2

Explicit moisture scheme (IMPHYS)

Grauple (gsfc)

Schultz

Non-FDDA

FDDA

Nudging

MM5 simulation was conducted by two nested domains with grid resolutions 27 km and 9 km. Domain 1 consisted of 35x41 grids covering the whole of Java island with 945x1107 km width. While, domain 2 consisted of 49x52 grids covering the 441x468 km width area comprising Central Java province. Figure 6 shows the geographical locations of domain 1 and 2.

12 12

170


Domain 2 Domain 2

Domain 1 Domain 1

Figure 6: Geographical location of domain 1 and domain 2 in MM5 simulation Figure 6: Geographical location of domain 1 anddomain domain 22 in Figure 6: Geographical location of domain 1 and in MM5 MM5simulation simulation The distribution of daily precipitation from TRMM, MM5, and MM5 with FDDA data that The of daily precipitation from from TRMM, MM5, MM5, and MM5 FDDA occured in the The distribution of time daily precipitation TRMM, andwith MM5 withdata data that occured indistribution the simulation periode during heavy precipitation event (January 5FDDA – that 7, 2005) and simulation time periode during heavy precipitation event (January 5 – 7, 2005) and low precipitation event (January occured in the simulation time periode heavy event (January 5 – 7,8.2005) and low precipitation event (January 23 – during 25, 2005) areprecipitation shown in Figure 7 and Figure In heavy 23 –precipitation 25, 2005)event, are shown Figure 723 and 8. In heavy precipitation event, it and is indicated that distribution low event (January –Figure 25, 2005) areofshown in Figure Figure 8. In heavy of precipitation it isin indicated that distribution precipitation is 7presented quite different. precipitation it quite isMM5 indicated that ofdata precipitation presented quite precipitation isevent, presented different. Gooddistribution relation between MM5 and TRMM data wasdifferent. shown in low Good relation between simulation and TRMM wassimulation shownisin low precipitation events. Good relation between MM5 simulation and TRMM data levels was in relatively low precipitation events. precipitation events. The precipitation pattern presents distribution which are both simulation The precipitation pattern presents distribution levels which are shown relatively equal inequal bothinsimulation The precipitation pattern presents distribution levels which are relatively equal in both simulation results and results andTRMM. TRMM. results and TRMM.

Figure 7: Daily precipitation distribution resulted by MM5 and TRMM

Figure 7: Daily during precipitation distribution MM5 peak precipitation eventresulted (Januaryby 5-7, 2005)and TRMM Figure 7: Daily precipitation distribution resulted by MM5 and TRMM


13

171

duringduring peak precipitation event (January 5-7, 2005) peak precipitation event (January 5-7, 2005)

Figure 8: Daily precipitation distribution resulted by MM5 and TRMM

FigureFigure 8: Daily distribution resulted byJanuary MM5 and TRMM during low precipitation event (23-25, 8: precipitation Daily precipitation distribution resulted by2005) MM5 and TRMM duringduring low precipitation event (23-25, January 2005) low precipitation event (23-25, January 2005) Figure 9 (a, b, c, d and e) shows precipitation distribution of MM5 simulation at each meteorological station in Figure 9 (a, c, db,and shows precipitation distribution of MM5 simulation at eachat each 9observation (a, c, de) and e)TRMM shows precipitation distribution of MM5 simulation comparison withFigure the fieldb, data and satellite.

meteorological stationstation in comparison with the field data and TRMM satellite. meteorological in comparison with theobservation field observation data and TRMM satellite.

(a): Semarang stationstation (a): Semarang

(b): Simongan stationstation (b): Simongan

(c): Gunungpati stationstation (c): Gunungpati

(d): Sumurjurang stationstation (d): Sumurjurang

14

172


(e): Ungaran station

Figure 9: Comparison of daily precipitation at each station

Figure 9: Comparison of daily precipitation at each station

These figures even showed some biases by comparing observed and re-analysis (TRMM) data, and simulated figures showed some bythese comparing and re-analysis (TRMM) and precipitation;These generally, trendeven of precipitation that biases occur at stations observed show a similar pattern. At Semarang data, and simulated precipitation; generally, trend of precipitation that occur at these stations show Simongan stations, simulated results are in a good relation with the observed data eventhough a large difference of total a similar pattern. At Semarang and Simongan stations, simulated results are in a good relation with precipitation was shown on January 1 to 8, 2005. MM5 results at these two stations were underestimated, compared the observed data eventhough a large difference of total precipitation was shown on January 1 to 8, to observed and TRMM data. On the contrary, differences were also encountered at Gunungpati, Sumurjurang, and 2005. MM5 results at these two stations were underestimated, compared to observed and TRMM Ungaran stations especially on January 5 to 8, 2005. In these three stations, MM5 results was underestimated, compared data. On the contrary, differences were also encountered at Gunungpati, Sumurjurang, and Ungaran to thestations observedespecially and TRMM ondata. January 5 to 8, 2005. In these three stations, MM5 results was underestimated, Total cumulative of precipitation for each data. meteorological station in simulation period (30 days) shows an extreme compared to the observed and TRMM deviation as depicted in Table 3. of precipitation for each meteorological station in simulation period (30 Total cumulative

days) shows an extreme deviation as depicted in Table 3.

Tabel 3: Deviation of MM5 simulation and observation data

Tabel 3: Deviation of MM5 simulation and observation data

Total Deviation Totalcumulative cumulativeofofprecipitation Precipitation(mm) (mm) Deviation Observation TRMM MM5 MM5 MM5 MM5 Observation TRMM MM5 TRMM TRMMMM5 MM5 StationsStations (mm) (FDDA) (FDDA) (mm) (mm) (mm) (mm) (mm) (FDDA) (mm)(mm) (mm) (mm) 241 282 31 53 -41 210 188 Semarang Semarang 241 282 31 53 -41 210 209 282 82 148 -73 127 61 Simongan Simongan 209 282 82 148 -73 127 301 105 322 -199 -3 -220 Gunungpati 102 102 Gunungpati 301 105 322 -199 -3 288 301 14 14 -123 -123 -136 Ungaran 165 165 Ungaran 151 151 288 301 170 367 137 137 268 Sumurjurang 438 438 Sumurjurang 301 301 170 367 26871

4.2 Hydrological modeling

MM5 (FDDA) 188 61 -220 -136 71

4.2 Hydrological In thismodeling study, hydrological simulation in the Garang watershed with several sensible parameters performedsimulation in 2005. inInthe order to verify HSPF results, daily discharge was In this study,was hydrological Garang watershed with severalsimulated sensible parameters was performed compared the HSPF observed daily discharge of Garang-Panjangan station. Comparison both of in 2005. In order with to verify results, simulated daily discharge was compared with the observed daily of discharge datasets has given results,ofasboth shown in Figure 10. good results, as shown in Figure 10. Garang-Panjangan station.good Comparison datasets has given


15 173

Figure 10:10: Comparison ofofdaily discharge between simulated and observed in the Figure Comparison daily discharge simulated and observed in the Figure 10: Comparison of daily discharge between between simulated and observed datadata indata the Garang-Panjangan station result ofHSPF HSPF simulation Garang-Panjangan station as aas of HSPF simulation Garang-Panjangan station asresult aa result of simulation

Statistic analysis resulted a coeffisient of correlation (R) of simulation model 0.89 and coefficient of determination Statistic analysis correlation(R) (R)ofofsimulation simulation model Statistic analysisresulted resultedaa coeffisient coeffisient of correlation model 0.890.89 and and 22 of 0.79. According to the United(R Environmental Protection EPA, 2007), for hydrological model, (R2) coefficient 0.79. According According to United States Environmental Protection determination )2 )ofof0.79. to the theAgency United(US States Environmental Protection coefficient of of determination (RStates 2 2 if R Agency range between 0.85 – 0.90 and R between 0.7 0.8, a good relation is being perfomed. The performance of this between EPA, 2007),for forhydrological hydrological model, model, if andand R R between Agency (US(US EPA, 2007), if RR range rangebetween between0.85 0.85– 0.90 – 0.90 simulation reflects a good computation jobperfomed. by the numerical model and theof results meetsreflects a reflects good criteria. The - 0.8, a good relation being perfomed. The performance simulation a good 0.7 0.7 - 0.8, a good relation isisbeing The performance ofthis thisalmost simulation a good Nash-Sutcliffe Efficiency (NSE) criteria show a value of 0.68. According to Moriasi et al. (2007), this simulation model computation job by the numerical model and the results almost meets a good criteria. The computation job by the numerical model and the results almost meets a good criteria. The Nash-Sutcliffe Efficiency (NSE) criteria aa value to to Moriasi et al. can be judged as satisfactory because NSE value showes isshowes higher than 0.5. of Nash-Sutcliffe Efficiency (NSE) criteria value of0.68. 0.68.According According Moriasi et (2007), al. (2007), this simulation model can be judged as satisfactory becausedaily NSEdischarge value is of higher than 0.5. Figure 11 presents scatter plots of observed and simulated the Garang River this simulation model can be judged as satisfactory because NSE value is higher than 0.5. at GarangFigure 11 presents scatter plots of observed and simulated daily discharge of the Garang PanjanganFigure station. 11 presents scatter plots of observed and simulated daily discharge of the Garang

River at Garang-Panjangan station.

River at Garang-Panjangan station.

Figure 11: Scatter plots of of observed dailydischarge discharge Garang River Figure 11: Scatter plots observedand andsimulated simulated daily in in thethe Garang River

Figure 11:calibration Scatter plots of observed and that simulated daily discharge the Garang River Based on all calibration parameters verified asit mentioned above, can be Based on all parameters that were verified aswere mentioned above, can beinconcluded that itthis simulation concluded that this simulation model produces the reasonable result and good capability in order to model produces the reasonable result and good capability in order to represent hydrological condition and behavior in represent condition and behavior that in thewere Garang watershed. Basedhydrological on all calibration parameters verified as mentioned above, it can be the Garang watershed.

concluded that this simulation model produces the reasonable result and good capability in order to represent hydrological condition and behavior in the Garang watershed.

174

16 Adi PRASETYO, Takao YAMASHITA

4.3 Atmospheric and hydrological interaction interaction 4.3 Atmospheric and hydrological The simulation within the atmosphere-land surface interaction one of the goals of this The meteorological The simulation within the atmosphere-land surfaceisinteraction is one of study. the goals of this study. The meteorological data from MM5 simulation result will be used to set up the hydrological data from MM5 simulation result will be used to set up the hydrological model (HSPF). Combining the input data from model Combining thethe input data from MM5 and the parameter settingstoofcreate the calibration MM5 and (HSPF). the parameter settings of calibration which was done by HSPF, it is possible a 100% simulated which was done by HSPF, it is possible to create a 100% simulated scenario. sendsprocess. scenario. MM5 sends precipitation and temperature to the HSPF hydrological model in a one wayMM5 interaction precipitation and temperature to the HSPF hydrological model in a one way interaction process. HSPF calculations is initiated, but without feed back to MM5. HSPF calculations is initiated, but without feed back to MM5. The meteorological data input for HSPF model is derived from the MM5 and MM5+FDDA results obtained from The meteorological data input for HSPF model is derived from the MM5 and MM5+FDDA January 1 –obtained 30, 2005.from Comparison of 2005. hydrological simulation for each source of datasimulation input can befor seen in Figure results Januaryresults 1 – 30, Comparison results of hydrological each 12.source of data input can be seen in Figure 12.

Figure 12: Comparison of observed and simulation daily discharge using different meteorological Figure 12: Comparison of observed and simulation daily discharge using different meteorological input data input data Table 4: Comparison of HSPF simulation results using station based and MM5 input Table 4: Comparison of HSPF simulation results using station based and MM5 input

Hydrology model

(R2)

Station Stationbased based(HSPF) (HSPF)

0.1790.179

MM5model model MM5

0.3420.342

Hydrology model

MM5+FDDA

MM5+FDDA

(R2)

0.292

0.292

According to Table 4, R2 value2 of MM5 and MM5-FDDA shows a higher value than the station based model According to Table 4, R value of MM5 and MM5-FDDA shows a higher value than the (HSPF). The R2 value for MM5 and MM5-FDDA, as input data, are 0.342 and 0.292. The interesting point is that high MM5 and MM5-FDDA, as input data, are 0.342 and station based model (HSPF). The R2 value for 2 improvement was produced by the MM5 model. R increased 0.163was higher than MM5+FDDA (0.113)model. if compared 0.292. The interesting point is that high improvement produced by the MM5 R2 with theincreased original station based model input (HSPF). 0.163 higher than MM5+FDDA (0.113) if compared with the original station based Based on the statistical value, simulation result which combines the MM5 meteorological model as input data and model input (HSPF). Based on simulation, the statistical value, simulation HSPF hydrological provides better results. result which combines the MM5 meteorological

model as input data and HSPF hydrological simulation, provides better results.

175


17

4.4 Coastal/Estuary modeling The coastal/estuary modeling in this study is attempted to be set up in the Garang estuary. This estuary is situated on the northward coast of Java island directly connected to the Java Sea. The model calibrated was used to conduct a 4.4 Coastal/Estuary modeling simulation of combined freshwater and tidal flows to investigate flow velocities, salinity structure and distribution of The coastal/estuary this study is attempted to be set up in the Garang estuary. sediment concentration as well asmodeling the estuaryin circulation in the Garang River system. This estuary is situated on the northward coast of Java island directly connected to 0the Java Sea. The computational domain of the Garang estuary as shown in Figure 13, is located at SL 6 55’ 35’’ - 60 57’ 31’’ The model calibrated0 was used to conduct a simulation of combined freshwater and tidal flows to andinvestigate EL 1100 23’flow 12’’ -velocities, 110 24’ 56”. The model computation for Garangofestuary has horizontal finite difference salinity structure and distribution sediment concentration as well asmeshes of the 120 estuary x 120 grids with 25 m interval for each grid (3 km x 3 km), vertically discritized with 10 σ-levels. The various circulation in the Garang River system. parameters The usedcomputational in the numericaldomain simulations are listed in Table 5 as follows: of the Garang estuary as shown in Figure 13, is located at SL 60

55’ 35’’ - 60 57’ 31’’ and EL 1100 23’ 12’’ - 1100 24’ 56”. The model computation for Garang 5: Parameters setting estuary has horizontal finite difference Table meshes of 120 x 120 grids with 25 m interval for each grid (3 km x 3 km), vertically discritized with 10 σ-levels. The various parameters used in the numerical X listed in Table 5 as follows: 25 m, 120 grids simulationsD are DY

25 m, 120 grids

Table 5: Parameters setting 1 s External interval time �X 25 m, 120 grids Internal interval time (DTI) 2s �Y 25 m, 120 grids NSTEPS 950400 External interval time 1s InternalSimulation interval time (DTI) 2 s 15, 2007 (22 days) time November 24 - December NSTEPS 950400 Initial Temperature 25 oC Simulation time November 24 - December 15, 2007 (22 days) o Initial Salinity 32 C Initial Temperature 25psu Freshwater salinity 0.02 psu Initial Salinity 32 psu Freshwater 0.02 psu Initialsalinity concentration of TSS 1 mg/l Initial concentration of TSS 1 mg/l Elevation boundary condition 300 tidal constituent Elevation boundary condition 300 tidal constituent 3 12.99 12.99mm/s3/s River inflow River inflow As initial input, divided into 3 grid cells = 4.33 m3/s 3 As initial input, divided into 3 grid cells = 4.33 m /s Cohesive sediment consentration 38.88 Cohesive sediment consentration 38.88kg/m kg/m3 3

X Y

Figure 13: Horizontal grid of the Garang estuary in ECOMSED model

Figure 13: Horizontal grid of the Garang estuary in ECOMSED model

18

Results of the hydrodynamic model were validated by comparing calculated tidal levels Adi PRASETYO, Takao YAMASHITA with measurement in Semarang and Tanjungmas Port (marked with X and Y point in Figure 14) duringResults the period of December 14-15,model 2007. were Figurevalidated 15 showsbythecomparing comparison between simulation of the hydrodynamic calculated tidal levels results and observation data for sea elevation. with measurement in Semarang andsurface Tanjungmas Port (marked with X and Y point in Figure 14) Results of the hydrodynamic model were validated by comparing calculated tidal levels with measurement in during the period of December 14-15, 15 shows thethe comparison between Semarang and Tanjungmas Port (marked with 2007. X and YFigure point in Figure 14) during period of December 14-15,simulation 2007. results and observation data for sea surface elevation. Figure 15 shows the comparison between simulation results and observation data for sea surface elevation. 176

Figure 15: Comparison of Sea Surface Elevation (SSE) between observed data and simulated result at Semarang and Tanjungmas Port Figure 15: Comparison of Sea Surface Elevation (SSE) between observed data and simulated

Figure 15: Comparison of Sea Surface (SSE) between result Elevation at Semarang and Tanjungmas Port2observed data and simulated result ) of simulation models compared According to Figure 16,atcoefficient determination Semarangofand Tanjungmas(R Port with Semarang station indicated 0.83 and 0.85 compared with Tanjungmas Port station. The According to Figure 16, coefficient of determination (R2) of simulation models compared with Semarang station simulation results 0.85 of sea surface elevation changes shows very good agreement with observed data indicated 0.83 and to compared withcoefficient Tanjungmas Port The simulation of sea surface elevation ) of simulation modelschanges compared According Figure 16, of station. determination (R2results of 48 hours simulation It also shows that the model seems be able to reproduce tidal propagation very goodstation agreement with observed dataand of 48 0.85 hours simulation It also shows that the model Port seems be able to The with shows Semarang indicated 0.83 compared with Tanjungmas station. wellreproduce in tide tidal and propagation river flowwell conditions. in tide and river flow conditions. simulation results of sea surface elevation changes shows very good agreement with observed data of 48 hours simulation It also shows that the model seems be able to reproduce tidal propagation well in tide and river flow conditions.

16: plots Scatterof plots of surface elevation between observed data and simulated result Figure 16:Figure Scatter surface elevation between observed data and simulated result Figure 17 the the simulation results of sediment of the date that observed available. Figure 17shows shows simulation results ofconcentration sediment concentration of the data dateare that observed These results show that distribution of suspended sediment in Garang estuary. During the time period of simulation, Figure 16: Scatter plots of surface elevation between observed data and simulated result data are available. These results show that distribution of suspended sediment in Garang estuary. it is indicated that sedimentation of simulated results located precisely at the end of the Garang estuary, produced During the time period of simulation, it is indicated that sedimentation of simulated results located approximately 32 mg/l – 40 mg/l of concentrates. If comparing these data with the observed data, it can be concluded Figure 17 shows results ofproduced sedimentapproximately concentration of date– that observed precisely at the end ofthethesimulation Garang estuary, 32 the mg/l 40 mg/l of that simulation results of November 24, 2007 (1st day) is in good performance (observed data = 38.8 mg/l), despite the data are available. These results that the distribution suspended in that Garang estuary. concentrates. If comparing these show data with observedofdata, it can besediment concluded simulation fact that simulation result of December 8, 2007 (15th day), is showing a large difference with simulated results being During time period24,of2007 simulation, that sedimentation simulated resultsdespite located good performance (observedofdata = 38.8 mg/l), results the of November (1st day)itisisinindicated much smaller than observed data (measured data was 242.67 mg/l). th

is showing32 a large withof the fact that simulation of December 8, 2007 (15 day), precisely at the end ofresult the Garang estuary, produced approximately mg/ldifference – 40 mg/l simulated results being much smaller data (measured data 242.67 mg/l). concentrates. If comparing these datathan withobserved the observed data, it can bewas concluded that simulation results of November 24, 2007 (1st day) is in good performance (observed data = 38.8 mg/l), despite the fact that simulation result of December 8, 2007 (15th day), is showing a large difference with simulated results being much smaller than observed data (measured data was 242.67 mg/l).


Conc.[mg/l]:

0

8

16

24

32

40

Conc.[mg/l]:

0

8

16

24

32

40

Conc.[mg/l]: 1days

0

8

16

24

32

40

Conc.[mg/l]: 15days 0

8

16

24

32

40

2.5

1days

2.5 2

2.5 2

2 1.5

2 1.5

Y[km]Y[km]

Y[km]Y[km]

2.5

1.5 1 1 0.5 0.5 0 0

19

177

15days

1.5 1 1 0.5

0

1

0

0.5 0

2

X[km]

1

0

2

0

1

0

1

2

X[km]

2

X[km] X[km] FigureFigure 17: Computation of horizontal sediment ECOMSED model 17: Computation of horizontal sedimentconcentration concentration in ECOMSED model Figure 17: Computation of horizontal sediment concentration in ECOMSED model Figure 18 shows the vertical profile of suspended sediment concentration the longitudinal Figure 18 shows the vertical profile of suspended sediment concentration in the in longitudinal section that section that represents the position of the Garang estuary and Java Sea. These figures were plotted Figure 18 shows the vertical profile of suspended sediment concentration in the longitudinal represents the position of the Garang estuary and Java Sea. These figures were plotted in order to visualize and interpret in orderthat to visualize and vertical transport. Thisand figure, clearlyfigures the distribution of section represents theinterpret position of thetheGarang estuary Javashows Sea. These were and plotted vertical transport. This figure, shows clearly distribution of suspended sediment between fresh salt water suspended sediment between freshvertical and salt water interface. It isshows also indicated that material of of in order to visualize and interpret transport. This figure, clearly the distribution interface. It is also indicated that material of sediment with concentration higher than 30 mg/L ends up at a distance of sediment with concentration higher than 30 mg/L ends up at a distance of 0.5 km from boundary suspended sediment between fresh and salt water interface. It is also indicated that material of 0.5border km from border near the Garang estuary.itThe closer into it empties into the the sea, the smaller or boundary precisely near or theprecisely Garangthan estuary. The ends closer the the sediment with concentration higher 30 mg/L up atempties a distance ofthe 0.5sea, km fromsmaller boundary sediment a rate below 5below mg/L.5 The sediment a rate estuary. mg/L. borderconcentration orconcentration preciselyshowing nearshowing the Garang closer it empties into the sea, the smaller the sediment concentration showing a rate below 5 mg/L.

1

0

5

10

15

20

25

30

35

Conc.[mg/l]:

0

5

10

15

20

25

30

35

1 1 0

0 -1

0 -1

-1 -2

-1 -2

-2 -3

1days

-3 -4

1days

Depth[m] Depth[m]

Depth[m] Depth[m]

1 0

Conc.[mg/l]:

-4 -5 -5 -6

0

5

10

15

20

25

30

35

Conc.[mg/l]:

0

5

10

15

20

25

30

35

-2 -3

15days

-3 -4

15days

-4 -5 -5 -6

-6 -7 -7

Conc.[mg/l]:

-6 -7 0

0.5

0

0.5

1

1.5

2

1

1.5

2

Distance[km]

-7

0

0.5

0

0.5

1

1.5

2

1

1.5

2

Distance[km]

Distance[km] Figure 18: Vertical distribution of sediment concentration resultDistance[km] in ECOMSED model FigureFigure 18: Vertical distribution of sediment concentration in ECOMSED ECOMSEDmodel model 18: Vertical distribution of sediment concentrationresult result in Sea surface temperature (SST) in and around the Garang estuary as one of some outputs of theSea ECOMSED modeling result,in was with re-analysis data theoutputs Oceanic surface temperature (SST) and compared around Garang estuary one from ofassome the ECOMSED Sea surface temperature (SST) in and the around the Garangas estuary one ofNational someofoutputs of and Atmospheric Adminstration (NOAA), Earth System Research Laboratory. The SST re-analysis the ECOMSED was compared withthe re-analysis data from National Oceanic modeling result, wasmodeling comparedoresult, with re-analysis data from National Oceanic andthe Atmospheric Adminstration FigureThe 19,Earth shows that the approximate temperature resulted from data is Earth weekly based on 1 grid. and Atmospheric Adminstration (NOAA), System Research Laboratory. The1oSST (NOAA), System Research Laboratory. SST re-analysis data is weekly based on grid.re-analysis Figure 19, shows o 24 oC – 25 oC, whereas output of NOAA re-analysis ranges higher, ECOMSED ranges between grid. Figure 19, shows that thebetween approximate temperature is weekly based on 1 resulted o thatdata the approximate temperature from ECOMSED ranges 24 oC – 25 C, whereasresulted output offrom NOAA reo o o C – 28.5 observed periods (Figure 20). namely between 27.5between output of NOAA re-analysis ranges higher, ECOMSED ranges 24 CoCat–all o25 C, whereas o analysis ranges higher, namely between o o 27.5 C – 28.5 C at all observed periods (Figure 20). namely between 27.5 C – 28.5 C at all observed periods (Figure 20).

178

20 20


Temp.[deg.]: Temp.[deg.]: 2.5 2.5

23 23

1days 1days

23.4 23.4

23.8 23.8

24.2 24.2

24.6 24.6

25 Temp.[deg.]: 25 Temp.[deg.]: 2.5 2.5

1.5 1.5

23.8 23.8

24.2 24.2

24.6 24.6

25 25

1.5 1.5

1 1

1 1

0.5 0.5

0.5 0.5

0 0 0 0

15days 15days

23.4 23.4

2 2

Y[km] Y[km]

Y[km] Y[km]

2 2

23 23

1 1

X[km] X[km]

2 2

0 0 0 0

1 1

X[km] X[km]

2 2

Figure 19: Computation of the horizontal distributionof of temperature temperature ininECOMSED model Figure 19: Computation of the horizontal distribution ECOMSED model Figure 19: Computation of the horizontal distribution of temperature in ECOMSED model

(a) week-1stst (November 18 – 24, 2007) (a) week-1 (November 18 – 24, 2007)

Temp oC

(b) week-4th (December 8 – 15, 2007)Temp oC (b) week-4th (December 8 – 15, 2007)

FigureFigure 20: Weekly sea surface temperature (SST) Indonesia basedon on NOAA re-analysis 20: Weekly sea surface temperature (SST)in Indonesia based re-analysis datadata Figure 20: Weekly sea surface temperature (SST) ininIndonesia based onNOAA NOAA re-analysis data According to Wyrtki (1961), average temperature in the Java Sea ranges fromo 27 ooC to 30 o to Wyrtki (1961), average in the Java in Seathe ranges C tofrom 30 C.27This C temperature to 30 According to Wyrtki (1961), temperature average temperature Javafrom Sea 27 ranges o According oC. This temperature is quite high for a typical tropical area range. There are some quite big

C. high Thisfor temperature is quite a typical tropical area range. between There are some quite big is quite abetween typical tropical area high range.for There are some quite big ECOMSED differences the ECOMSED simulation results anddifferences re-analysis data the from NOAA simulation and differences between the ECOMSED simulation results and re-analysis data size from NOAA area and results and re-analysis data from NOAA and Wyrtki, which are estimated due to the small of domain Wyrtki, which are estimated due to the small size of domain area of the model which covers onlyof the Wyrtki, which are estimated due to the small size of domain area of the model which covers only model width areadomain of 3 x area 3 km.isIfenlarged, the domain is enlarged, sea surface temperature the which width covers area ofonly 3 x the 3 km. If the seaarea surface temperature may possibly be may the width area of 3 x 3 km. If the domain area is enlarged, sea surface temperature may possibly be possibly to or precisely NOAA or Wyrtki research results. closerbetocloser or precisely equalsequals NOAA or Wyrtki research results. closer to or precisely equals NOAA or Wyrtki research results. 21 shows the simulation of surface theperiod time period of simulation. FigureFigure 21 shows the simulation resultsresults of surface salinitysalinity in the in time of simulation. The salinity Figure 21 shows the simulation results of surface salinity in the time period of simulation. The salinity distributions present the consistent pattern over the whole estuary. These figures distributions present the consistent pattern over the whole estuary. These figures indicate some influences of river The salinity distributions present the consistent pattern over the whole estuary. These figures indicate some influences of river outflow near the estuary as indicated by the salinity rate in this outflow near some the estuary as indicated by the salinity ratethe in this area as which ranges between 0 to 10 rate psu exactly indicate influences of river outflow near estuary indicated by the salinity in this at the area which ranges between 0 to 10closer psu exactly at higher the confluence of fresh and salt water. 30 The closer confluence of fresh and between salt water.0 The to sea, the salinity range fromand approximately to 32 psu. area which ranges to 10 psu exactly at the confluence of fresh salt water. Thepsu closer to sea, the higher salinity range from approximately 30 psu to 32 psu. According to Veen (1953), it to sea, to theVeen higher salinity range that fromtheapproximately psu tovaries 32 psu. According to 30.9 Veenpsu (1953), According (1953), it is stated average surface30 salinity annually between to 34.3itpsu in is stated that the average surface salinity varies annually between 30.9 psu to 34.3 psu in the stated that theJava average salinity varies annually 30.9 psu toto32.6 34.3 psu in the the is eastern part SeaSea andsurface becoming smaller to westward withbetween a with rangeaofrange 30,6 psu eastern partofofthe the Java and becoming smaller to westward of 30,6 psupsu. to 32.6 psu. eastern part of the Java Sea and becoming smaller to westward with a range of 30,6 psu to 32.6 psu.

21 21 Numerical Study on Atmospheric, Hydrological and Coastal Circulation in Coastal City, Semarang Indonesia

Sal.[psu]: Sal.[psu]:

0 0

5 5

10 10

15 15

20 20

25 25

30 30

35 35


5 5

10 10

15 15

20 20

25 25

30 30

35 35

15days 15days

2.5 2.5

2.5 2.5

2 2

2 2

Y[km] Y[km]

Y[km] Y[km]

1days 1days

0 0

179

1.5 1.5

1.5 1.5

1 1

1 1

0.5 0.5

0.5 0.5

0 00 0

1 1

0 00 0

2 2

X[km] X[km]

1 1

2 2

X[km] X[km]

FigureFigure 21: Computation of the distribution ECOMSEDmodel model 21: Computation of horizontal the horizontal distributionofofsalinity salinity in in ECOMSED Figure 22 shows the vertical salinity. These plotted in order Figure 22 shows the vertical profile ofprofile salinity.ofThese figures werefigures plotted were in order to visualize and to interpret visualize and interpret vertical transport of salinity in the model area. From these figures, can be vertical transport of salinity in the model area. From these figures, can be concluded that vertical salinity distribution concluded that vertical salinity distribution change at the sea region is not extremely large with a change at the sea region is not extremely large with a range of 30 psu to 32 psu. This conclusion is also supported by range of 30 psu to 32 psu. This conclusion is also supported by the statement of Wyrtki (1961) who thedeclared statementthat of Wyrtki (1961) who declared vertical profiles salinitysuch in shallow sea region, such as in generally vertical profilesthat ofgenerally salinity in shallow sea ofregion, as in the Java Sea, theshow Java Sea, show an almost homogeneous salinity from surface to bottom. an almost homogeneous salinity from surface to bottom. Sal.[psu]: Sal.[psu]:

0 0

7 7

14 14

21 21

28 28

35 35

0 0

0 0

-1 -1

-1 -1

-2 -2

-2 -2

1days 1days

-3 -3


1 1

Depth[m] Depth[m]

Depth[m] Depth[m]

1 1

-4 -4

7 7

14 14

21 21

28 28

35 35

15days 15days

-3 -3 -4 -4

-5 -5

-5 -5

-6 -6

-6 -6

-7 -7

0 0

-7 -7 0 0

0.5 0.5

1 1

Distance[km] Distance[km]

1.5 1.5

2 2

0 0

0.5 0.5

1 1

Distance[km] Distance[km]

1.5 1.5

2 2

22: Computation of the vertical distributionofofsalinity salinity in in ECOMSED FigureFigure 22: Computation of the vertical distribution ECOMSEDmodel model

6. 6. Conclusion Conclusion In this different modelsmodels as partasofpart the Regional Environment SimulatorSimulator (RES) such as thesuch Atmospheric In study, this study, different of the Regional Environment (RES) as model (MM5), Hydrological simulationHydrological programs (HSPF) and Estuary/Coastal model and (ECOMSED) were applied. the Atmospheric model (MM5), simulation programs (HSPF) Estuary/Coastal (ECOMSED) were applied.ofFor the MM5 model in 30, Formodel atmospheric modeling, simulation the atmospheric MM5 model modeling, in Semarangsimulation was done of in January 1 until January Semarang was done in precipitation January 1 until 2005. of daily from 2005. Distribution of daily fromJanuary TRMM,30, MM5, andDistribution MM5 with FDDA dataprecipitation have been compared. In TRMM, MM5, and MM5 with FDDA data have been compared. In general, eventhough biases general, even though biases were found when comparing observed data, re-analysis (TRMM) data and simulated precipitation, trend of precipitation that occured at the stations has shown a similar pattern. Moreover, the MM5 model

180


is recommended as meteorological model to provide high resolution of gridded climatologic data as closely as possible with the observation data. In hydrology modeling, HSPF model simulation in the Garang watershed with several sensible parameters was performed from January 1 to December 31, 2005. In order to verify the HSPF results, simulated daily discharge was compared with the observed daily discharge at Garang-Panjangan station. The comparison between both data gave a good results with correlation coefficient (R) of 0.89, determination coefficient (R2) of 0.79 and Nash-Sutcliffe Efficiency (NSE) criteria of 0.68. In overall, performance of simulation in numerical model has reflected a good computation and the result has met a good criteria. It can be concluded that the simulation model produced a reasonable result and good capability to represent the hydrological condition and behavior in the Garang watershed. The model computation for coastal estuary modeling was performed between November 24 and December 15, 2007. Whereas, model validation was carried out by comparing calculated tidal levels with measurement of Semarang and Tanjungmas Port and simulated results during the period of December 14 – 15, 2007. The comparison of simulated result of sea surface elevation changes for 48 hours simulation shows good agreement with the observed data with determination coefficient (R2) of 0.83 if compared with the Semarang station, and 0.85 compared with the Tanjungmas Port station The results also provided a time series of temperature, salinity, and distribution of sediment concentration in the Garang estuary.

References Ashok, K., Z. Guan, and T. Yamagata. (2001) Impact of the Indian Ocean Dipole on the relationship between Indian Ocean monsoon rainfall and ENSO. Geophysical Research Letter 28:4, p.499- 502. Bicknel, B.R., J.C Imhoff, J.L. Kittle, T.H. Jobes, and A.S. Donigian Jr., (2001) Hydrological Simulation Program Fortran: User’s Manual for Version 12: U.S. Environmental Protenction Agency, Athens, Ga., EPA/600/ R-97/080. Blumberg, A. F. and G. L. Mellor. (1987) A description of a three-dimensional coastal ocean circulation model, in Three-dimensional coastal ocean models, coastal and estuarine sciences, Vol. 4, Editor N. Heap, American Geophysical Union, Washington, D.C., p.1-16. BPS Jawa Tergah (2005), Central Java in Figures (Jawa Tengah dalam angka) 2005. Semarang. Statistics Indonesia (Badan Pusat Statistik). Climate Change, The Fourth Assessment Report (AR4) of the United Nations Intergovernmental Panel on Climate Change (IPCC) 2007. Switzerland. The IPCC Secretariat. Grell, G., Dudhia, J. And Stauffer, D. R. (1994) A description of the fifth generation Penn State/NCAR mesoscale model (MM5), NCAR/TN-398+STR, p.117. Groot, R., Kwakman, P. and Velden, N. (2007) Masterplan Semarang 2025. TU Delft, the Netherland. Haggag, M. (2009) Regional Environment Simulator: Atmosphere Ocean Land Surface Hydrology Coupled Model. Dr. dissertation, Graduate School for International Development and Cooperation, Hiroshima University, Japan. Haggag, M. and Yamashita, T. (2009) Environmental simulator and its application to the analysis of the tropical cyclone Gonu in 2007, Journal of International Cooperation and Development, (in press) Vol. 15. Haggag, M., Yamashita, T., Lee, H. S. and Kim, K. (2008) A coupled atmosphere and multi layer land surface model for improving heavy rainfall simulation, Hydrol. Earth System Science Discuss. Vol. 5, p.1067-1100. HydroQual. (2002). A primer of ECOMSED version 1.3: Users Manual. HydroQual, Inc., Mahwah, New Jersey. USA. Kuroki, A. J. P. (2007) Hydrological analysis of water quality constituents transport and transform in Haji Dam watershed. Dr. dissertation, Graduate School for International Development and Cooperation, Hiroshima


181

University, Japan. Lee, H. S. (2008) Guide for High Performance Linux Cluster (HPLC). Graduate School for International Development and Cooperation, Hiroshima University. Marfai, M. A. (2003). GIS modelling of river and tidal flood hazards in a waterfront city: case study, Semarang City, Central Java, Indonesia. M.Sc. thesis, International Institute for Geo-Information and Earth Observation, ITC, Enschede, The Netherlands. Mellor, G. L. and Yamada, T. (1982) Development of a turbulence closure model for geophysical fluid problems. Rev. Geophys. Space Phys., 20, p. 851-875. Nash, J. E. and J. V. Sutcliffe. (1970) River flow forecasting through conceptual models: Part 1. A discussion of principles. J. Hydrology 10 (3): p.282-290. Petit, D., Cotel, P. and Nugroho, D. (1996) The seasonal variation of salinity in The Java sea. Procceding of Acoustics, p.29-41. Research Centre for Water Resources, Ministry of Public Works. Hydrology data. Bandung. Indonesia Sadhotomo, B. and Durand, J. R. (1996) General features of Java sea ecology. Procceding of Acoustics, p.43-54. Santhi, C. J. G., Arnold, J. R., William, W., A Dugas., R. Srinivasan. and L. M. Hauck. (2001) Validation of the SWAT model on a large river basin with point and nonpoint sources, J. American Water Resources Assoc. 37 (5): p.1169-1188. Semarang Drainage Masterplan Document. (2007) Semarang: Indonesia. Shrestha, P.L., A.F. Blumberg, D.M. Di Toro and F. Hellweger. (2000) A three-dimensional model for cohesive sediment transport in shallow bays, Invited Paper, ASCE Joint Conference on Water Resources Engineering and Water Resources. Planning and Management, July 30-August 2, 2000, Minneapolis, MN. The Agency for the Assessment and Application Technology (BPPT). (2006) Jakarta: Indonesia. U.S. EPA. (2000) BASINS Techinal Note 6: Estimating Hydrology and Hydraulic Parameters for HSPF. United States Environmental Protection Agency. EPA-823-R00-012 Van Rijn, L. C. (1993) Principles of sediment transport in rivers, estuaries and coastal seas, Aqua Publications, Amsterdam, the Netherlands. Veen, P .C. (1953) Preliminary charts of the means salinity of the Indonesia archipelago and adjacent waters, Org. Sci. Res. Indonesia, p.17-46. Wahyu, A. (2009) Annual and seasonal discharge responses to forest/land cover changes and climate variations in Kapuas river basin. M.Sc. thesis, Graduate School for International Development and Cooperation, Hiroshima University, Japan. Wyrtki, K. (1961) Scientific results of marine investigations of the South China Sea and the Gulf of Thailand 19591961. Naga Report, 2. University of California, Scripps Institute of Oceanography, La Jolla, California. Yamashita, T., Kim, K. O., Lee, H. S. and Haggag, M. (2007) Environment Simulator. Annual Journal of Coastal Engineering, JSCE, 54, p.1301-1305.